On the order of a generalized hexagon admitting an ovoid or spread (Q1604578)

This note contains a very unexpected result which follows from an almost trivial (hence very clever!) observation. A generalized hexagon of order \((s,t)\) is a point line geometry the incidence graph of which has girth 12 and diameter 6, and such that each line (point) is incident with exactly \(s+1\) points \((t+1\) lines). An ovoid \(O\) of a generalized hexagon is a set of points such that each point of the hexagon is collinear with exactly one point of \(O\) (collinearity includes equality here). A spread is the dual notion of an ovoid. If \(s\) and \(t\) are finite, then one can count the number of points of a potential ovoid. In the literature, one used different ways to count, but apparently nobody ever compared the results of these different values, until Alan Offer did in the paper under review. The theorem that emerges says that no finite generalized hexagon admits an ovoid unless \(s=t\). Unfortunately, the order of the known hexagons are of the form \((s,s)\), \((s,s^3)\), or \((t^3,t)\), and it has been proved before that in the latter two cases an ovoid cannot exist (using a divisibility condition). Still, the paper under review provides a far more elegant proof of this, with also a possibility for generalization. In joint work with the reviewer, some more nonexistence results of ovoids of finite generalized polygons have been obtained and will be published.